Re: roller chain instead of screws or rack?

Hello Dean, List,

You probably saw a Camtech CNC router... They sold HUNDREDS of
these "sprocket" drive machines(they didn't want to have to fight the
kind of responses you've gotten so far from the list!). Chain works
fine, and as your friend noted, for a long time as well.

When others switched to rack/pinion(digital tool, mostly;
they "really" started the "low cost cnc router" thing, IMO), Camtech
followed suit. I talked to one of their techs when I was about to
purchase a used camtech machine, since I had seen that they had
changed to rack/pinion. I asked him if the chain was being replaced
as an improvement for accuracy, or repeatability. HE SAID "NO!, IT'S
(rack/pinion) JUST WHAT THE MARKET WANTS!!" Specs are the SAME for
both the old chain drive and rack/pinion + - .004 .(Real world. cnc
marketing departments are terrible liars; at least in the under
$20K"router" arena)

The Camtech chain is #25, and using the aircraft cable removes some
cost and mass from the system. The ratios I've seen are 4:1, and use
a SS,34 motor with unipolar drives. The software was originally by
AXYZ (ak-seez), who now have their own complete line of machines.

The trick to using chain effectively is to distribute any error
across the entire range of travel. This is done by programming, then
measuring, a known distance. The difference between distance traveled
and distance SUPPOSED to have been traveled is used to figure the
TRUE steps per inch. Often, this will NOT be an integer; and software
that allows only integer step size will introduce errors. Note that
this is ALSO true for other types of drive systems! Chain works
REALLY well for "dirty" environments, unlike rack/pinion.

The only REAL "problem" with chain drive is that it "gallops", but
the cut quality of the Camtech machine(s) show that this is not
really a problem. Coarse timing belt drives(as described in one
earlier response) will also gallop.

(added on proofreading pass) This galloping will also introduce small
errors in the DISTANCE traveled, but please see my P.S. below...

The Camtech machines use an output sprocket of 12T. Let's think of
this as a 12 sided polygon (dodecagon), and compare it to a circle.
We'll wrap some string around the two and begin rotating them to pull
the string in. You can see that the pull will be constant with the
circle, but will vary slightly with the "dodec", since the radius is
changing. (to visualise this further, imagine a 4,3, or even 2 sided
sprocket or pulley).

So the bottom line, IMO, is go for it! As long as you are not looking
for .001 type parts(and with plasma, wood, plastic, and even many
metal products, you won't be), chain WILL work FINE!

Hope this helps.

Ballendo

P.S. CNC "specsmanship" is a pet peeve. So many are claiming
ridiculous numbers for resolution, repeatability, and accuracy...
Often times the STRUCTURE OF THE MACHINE is moving FAR BEYOND what
these claims can support! And the worst part is that the "newbies"
start to BELIEVE they NEED this type of performance. And that they
CAN ACHIEVE IT! Well folks, it just aint so, IMO.

BTW, resolution is the EASIEST "number" to attain, and probably the
most misused. Repeatability is where it's at on an inexpensive
machine, and REAL accuracy just WON'T be there! But, a "repeatable"
machine CAN be "compensated" into producing highly accurate parts.

--- In CAD_CAM_EDM_DRO@y..., deanc500@y... wrote:
> Hi, I have built several small router tables out of both wood and
> metal with varying degrees of success. I used lead screws and
> anti-backlash nuts on all of them. I am now building a 5'x10'
plasma
> cutter. At first I thought rack and pinion was the best option, but
> then while recently visiting a sign shop for supplies, I noticed a
> large CNC router in the back. I asked to have a closer look and was
> surprised to see a #35 roller chain drive. <snip>